Resumen de web cluster dynamic load balancing: GA approach

Chin Wen Cheong, Amy Lim Hui Lan

• In this paper, genetic based Generalized Dimension Exchange (GDE) method is proposed with the intention to uniformly distribute the unprecedented web clusters workload. Arbitrary topology of web servers is paired based on the graph coloring algorithms. Combination of GDE and genetic algorithm robust searching capability will be used to explore the variety of exchange parameters to achieve the minimum convergence rate of the web cluster. Genetic algorithm (GA) based on the evolutionary concept will choose the fittest survived genome to optimize the distribution process. Simulation of 10 samples with different initial workload has been carried out in a 3x3 mesh web cluster.

  The proposed GA GDE reviewed a better stability measurement compare with the traditional GDE method.